From owner-freebsd-isp Sun Sep 10 12:13: 0 2000 Delivered-To: freebsd-isp@freebsd.org Received: from verdi.nethelp.no (verdi.nethelp.no [158.36.41.162]) by hub.freebsd.org (Postfix) with SMTP id 9289037B424 for ; Sun, 10 Sep 2000 12:12:57 -0700 (PDT) Received: (qmail 15370 invoked by uid 1001); 10 Sep 2000 19:12:53 +0000 (GMT) To: mike@mikesweb.com Cc: wizard@sybaweb.co.za, freebsd-isp@FreeBSD.ORG Subject: Re: NIC settings From: sthaug@nethelp.no In-Reply-To: Your message of "Sun, 10 Sep 2000 15:07:22 -0400" References: <4.3.2.7.2.20000910150718.00b3b530@mail.mikesweb.com> X-Mailer: Mew version 1.05+ on Emacs 19.34.2 Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Date: Sun, 10 Sep 2000 21:12:53 +0200 Message-ID: <15368.968613173@verdi.nethelp.no> Sender: owner-freebsd-isp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Actually, switching to half duplex won't really help you a whole lot.. I > had a 10mb hub that had the collision light almost always on. If you want > to stop the collisions you'll want to replace your hub with a switch. Why on earth do you want to stop the collisions? Collisions are *normal* and *expected* when you use half duplex Ethernet. Collisions seldom matter, performance does. As Rich Seifert puts it, > A major preoccupation with network administrators these days seems to be > monitoring and worrying about the number of collisions seen on Ethernet > networks. There is a great deal of folklore and voodoo concerning what > is an "acceptable" collision rate or collision percentage, and when is > the network "broken" or on the verge of collapse. Except in the most > extreme of circumstances (all of which are observable through other, > better metrics), the number of collisions seen on a network in an > uninteresting and misleading statistic. (If you don't know who Rich Seifert is, check out comp.dcom.lans.ethernet.) Steinar Haug, Nethelp consulting, sthaug@nethelp.no To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message